OutSystemsPerspectives

Addressing the Last Inefficient Industry – IT!

“Only 32% of IT projects succeed”

This got everyone’s attention as Paulo Rosado, CEO of OutSystems, began his NextStep keynote describing how IT departments are typically viewed by their business counterparts; in a word – inefficient.

He was showing a chart by Standish Group to a packed auditorium…that’s a pretty awful stat – but why this inefficiency? 

pauloday1.jpg

Paulo discussed the faults of the waterfall approach and how business change quickly outstrips IT’s ability to deliver and, therefore, the rapid decline into misalignment of IT and the Business. IT backlogs pile up and they don’t have enough time to respond to new business needs. To add to this, software agesaccording to Gartner, enterprise software reaches the end of its life every 5-7 years.

All of this, Paulo asserted, results in the cost of change rising exponentially, both as time passes and as the volume of required change increases.

Paulo shared how OutSystems and its customers and partners are seeing something different from the norm – Efficiency.

Here are some quotes he shared on how people view OutSystems-driven IT depts:

“IT is very responsive. If it is important, we get a fix in 1 week.”“Business units now come to us asking to have the project done with that ‘OutSystems thing.”

He shared a bunch of metrics from the OutByNumbers program which benchmarks OutSystems-based IT depts against each other and against traditional IT depts.  The net net of the slides he shared (which we will upload shortly) – is that in the OutSystems world, the cost of change is both low and stable.cost of change1.jpg

What makes this possible?  The Agile method, in combination with a technology that enables rapid and continuous change.

Drilling more deeply into the reasons for the rising cost of change in traditional approaches, Paulo compared them to the OutSystems approach and explained how we have been able to achieve the opposite of industry norms – and effectively addressing the last inefficient industry.

Four key areas of discussion:

1. Software Interdependency – and the domino effect of change in highly coupled architectures. The TrueChange engine in the Agile Platform alleviates the need to manually check for, and resolve the impact of change.

2. Monolithic Architecture that keep growing and don’t allow for reuse. The answer, Paulo asserts, is to create and reuse services as needs evolve and to ensure the software is broken up into smaller piece, that are connected through protocols: supported in the Agile Platform through refactoring and the support for iterative and bottom-up SOA.

3. Broken Application Lifecycles – where the connection between requirements, development, deployment and monitoring is not seamless. These are addressed in the Agile Platform by Embedded Change Technology, 1-click Publishing and Performance Management.

4. Low Knowledge Transfer – when developers leave the team, so does knowledge of the systems they worked on and it takes time to ramp new people. The Agile Platform answer: Model Driven Development.  Here Paulo shared an interesting example to show how models can simplify knowledge transfer: in a real customer’s environment 13,157,961 lines of code are represented in the model by 1602 processes, 987 web pages and 745 data structures.  The math is easy: the less elements there are to understand, the easier the knowledge transfer.

Paulo discussed the typical push-back he hears about the model-driven approach: it’s only for building toy apps vs. enterprise-class, and problems scaling to real business needs. He looked into the issues of traditional RAD tools, mashup platforms and the combination of the cloud.  Wrapping up with how the Agile Platform was built for scale and to enable the smooth transition from tactical solutions to strategic, enterprise-class.

He closed by saying you could stabilize the cost of software change at a low level, and make IT more efficient by doing the following:

1. Keep architecture fresh + 2. Automate change impact + 3.Address the full software lifecycle + 4. Simplify knowledge transfer.

 

About the author

Michel Ozzello

Although he's been working in Marketing for the past decade or so, Michel is still a geeky Software Engineer at heart. He tries to fit technology in every bit of marketing activity he does - from SEO to websites and digital advertising.

Comments

Michel

This was a great talk, and the number Paulo presented were impressive. Here are some notes I took during the session:
– One OutSystems customer went from 0 to 10K function points in 1 year (they built 9 business apps equivalent to a small ERP);
– They delivered 172 release (that 1 new release every 2 days) with an average of 232 Function Points per release;
43% of the changes for those releases were related to database changes (imagine the implications on the other layers, especially UI)
– With Model Driven Development, you basically transform 13M lines of code into ~3000 modeled elements. This makes knowledge transfer between developers possible

Michel

Here are some quotes I took from the keynote.
“1st Agile Platform version was built for Telecoms, that’s why it is prepared to scale”
“Everyone remembers the 90s Mushroom nightmare of Visual basic and Access apps spouting everywhere”
“Business users post feedback on the application, straight to IT, using Embedded Change Technology”
“Upfront SOA is too expensive. Re-factor towards SOA as needed!”
“From tactical to strategic, smoothly”

Leave Your Comment